Update app.py

app.py

@@ -1,5 +1,5 @@
 # app.py
-# A Flask API with extra print statements for debugging on Hugging Face Spaces.
+# A Flask API updated to use the proper Flask logger instead of print().
 
 import os
 import pandas as pd
@@ -8,31 +8,35 @@ from flask import Flask, request, jsonify
 from flask_cors import CORS
 from huggingface_hub import HfApi
 from datetime import datetime
+import logging
 
 # --- Initialization ---
-print("--- Flask app.py is starting up! ---")
 app = Flask(__name__)
 CORS(app)
 
+# Configure logging
+app.logger.setLevel(logging.INFO)
+app.logger.info("--- Flask app.py is starting up! ---")
+
 # --- Configuration ---
-print("Loading environment variables...")
+app.logger.info("Loading environment variables...")
 try:
     HF_TOKEN = os.environ["HF_TOKEN"]
     AUDIO_DATASET_REPO_ID = os.environ["AUDIO_DATASET_REPO_ID"]
     TEXT_DATASET_REPO_ID = os.environ["TEXT_DATASET_REPO_ID"]
-    print("Successfully loaded all required environment variables.")
+    app.logger.info("Successfully loaded all required environment variables.")
 except KeyError as e:
-    print(f"FATAL ERROR: Missing secret environment variable: {e}")
+    app.logger.error(f"FATAL ERROR: Missing secret environment variable: {e}")
     HF_TOKEN, AUDIO_DATASET_REPO_ID, TEXT_DATASET_REPO_ID = None, None, None
 
 # --- Hugging Face API Client ---
 if HF_TOKEN:
-    print("Initializing HfApi client...")
+    app.logger.info("Initializing HfApi client...")
     api = HfApi(token=HF_TOKEN)
-    print("HfApi client initialized.")
+    app.logger.info("HfApi client initialized.")
 else:
     api = None
-    print("Warning: HfApi not initialized because HF_TOKEN is not set.")
+    app.logger.warning("Warning: HfApi not initialized because HF_TOKEN is not set.")
 
 
 # --- Helper Function ---
@@ -46,7 +50,7 @@ def get_unique_filename():
 @app.route('/')
 def index():
     """A simple index route to confirm the API is running."""
-    print("Request received for / route.")
+    app.logger.info("Request received for / route.")
     return "Hugging Face Data Uploader API is running."
 
 @app.route('/add-audio', methods=['POST'])
@@ -54,24 +58,24 @@ def add_audio_data():
     """
     Receives an audio/transcription pair, converts to Parquet, and uploads to the audio dataset.
     """
-    print("Request received for /add-audio route.")
+    app.logger.info("Request received for /add-audio route.")
     if not api:
-        print("Error: API client not available.")
+        app.logger.error("API client not available for /add-audio.")
         return jsonify({"error": "Server is not configured with Hugging Face credentials."}), 500
     try:
-        print("Attempting to process /add-audio data...")
+        app.logger.info("Attempting to process /add-audio data...")
         data = request.get_json()
         if not data or 'audio' not in data or 'transcription' not in data:
-            print("Invalid payload received for /add-audio.")
+            app.logger.warning("Invalid payload received for /add-audio.")
             return jsonify({"error": "Invalid payload. 'audio' and 'transcription' fields are required."}), 400
         
-        print("Data validated. Creating DataFrame.")
+        app.logger.info("Data validated. Creating DataFrame.")
         df = pd.DataFrame([data])
         buffer = io.BytesIO()
         df.to_parquet(buffer, index=False, engine='pyarrow')
         buffer.seek(0)
         
-        print(f"Uploading file to audio dataset: {AUDIO_DATASET_REPO_ID}")
+        app.logger.info(f"Uploading file to audio dataset: {AUDIO_DATASET_REPO_ID}")
         api.upload_file(
             path_or_fileobj=buffer,
             path_in_repo=get_unique_filename(),
@@ -79,11 +83,11 @@ def add_audio_data():
             repo_type="dataset",
             commit_message="Add new audio-transcription pair"
         )
-        print("File successfully uploaded to audio dataset.")
+        app.logger.info("File successfully uploaded to audio dataset.")
         return jsonify({"message": "Audio data added successfully."}), 201
 
     except Exception as e:
-        print(f"---! UNEXPECTED ERROR in /add-audio !---: {e}")
+        app.logger.error(f"---! UNEXPECTED ERROR in /add-audio !---: {e}", exc_info=True)
         return jsonify({"error": "An internal error occurred.", "details": str(e)}), 500
 
 
@@ -92,24 +96,24 @@ def add_text_data():
     """
     Receives a transcription/summary pair, converts to Parquet, and uploads to the text dataset.
     """
-    print("Request received for /add-text route.")
+    app.logger.info("Request received for /add-text route.")
     if not api:
-        print("Error: API client not available.")
+        app.logger.error("API client not available for /add-text.")
         return jsonify({"error": "Server is not configured with Hugging Face credentials."}), 500
     try:
-        print("Attempting to process /add-text data...")
+        app.logger.info("Attempting to process /add-text data...")
         data = request.get_json()
         if not data or 'transcription' not in data or 'summary' not in data:
-            print("Invalid payload received for /add-text.")
+            app.logger.warning("Invalid payload received for /add-text.")
             return jsonify({"error": "Invalid payload. 'transcription' and 'summary' fields are required."}), 400
 
-        print("Data validated. Creating DataFrame.")
+        app.logger.info("Data validated. Creating DataFrame.")
         df = pd.DataFrame([data])
         buffer = io.BytesIO()
         df.to_parquet(buffer, index=False, engine='pyarrow')
         buffer.seek(0)
 
-        print(f"Uploading file to text dataset: {TEXT_DATASET_REPO_ID}")
+        app.logger.info(f"Uploading file to text dataset: {TEXT_DATASET_REPO_ID}")
         api.upload_file(
             path_or_fileobj=buffer,
             path_in_repo=get_unique_filename(),
@@ -117,14 +121,14 @@ def add_text_data():
             repo_type="dataset",
             commit_message="Add new transcription-summary pair"
         )
-        print("File successfully uploaded to text dataset.")
+        app.logger.info("File successfully uploaded to text dataset.")
         return jsonify({"message": "Text data added successfully."}), 201
         
     except Exception as e:
-        print(f"---! UNEXPECTED ERROR in /add-text !---: {e}")
+        app.logger.error(f"---! UNEXPECTED ERROR in /add-text !---: {e}", exc_info=True)
         return jsonify({"error": "An internal error occurred.", "details": str(e)}), 500
 
 # To run on Hugging Face Spaces
 if __name__ == '__main__':
-    print("Starting Flask development server...")
+    app.logger.info("Starting Flask development server...")
     app.run(host='0.0.0.0', port=7860)